A Permutation Importance-Based Feature Selection Method for Short-Term Electricity Load Forecasting Using Random Forest

The prediction accuracy of short-term load forecast (STLF) depends on prediction model choice and feature selection result. In this paper, a novel random forest (RF)-based feature selection method for STLF is proposed. First, 243 related features were extracted from historical load data and the time information of prediction points to form the original feature set. Subsequently, the original feature set was used to train an RF as the original model. After the training process, the prediction error of the original model on the test set was recorded and the permutation importance (PI) value of each feature was obtained. Then, an improved sequential backward search method was used to select the optimal forecasting feature subset based on the PI value of each feature. Finally, the optimal forecasting feature subset was used to train a new RF model as the final prediction model. Experiments showed that the prediction accuracy of RF trained by the optimal forecasting feature subset was higher than that of the original model and comparative models based on support vector regression and artificial neural network.

[1]  Jianjun Wang,et al.  An annual load forecasting model based on support vector regression with differential evolution algorithm , 2012 .

[2]  Jianzhou Wang,et al.  Short-term load forecasting using a kernel-based support vector regression combination model , 2014 .

[3]  Francisco Martinez Alvarez,et al.  Energy Time Series Forecasting Based on Pattern Sequence Similarity , 2011, IEEE Transactions on Knowledge and Data Engineering.

[4]  Sancho Salcedo-Sanz,et al.  Local models-based regression trees for very short-term wind speed prediction , 2015 .

[5]  Grzegorz Dudek,et al.  Pattern similarity-based methods for short-term load forecasting - Part 2: Models , 2015, Appl. Soft Comput..

[6]  Feng Yu,et al.  A short-term load forecasting model of natural gas based on optimized genetic algorithm and improved BP neural network , 2014 .

[7]  A. Arabali,et al.  A hybrid short-term load forecasting with a new input selection framework , 2015 .

[8]  Jaime Lloret,et al.  Artificial neural networks for short-term load forecasting in microgrids environment , 2014 .

[9]  Michael C. Fairhurst,et al.  Diversity in multiple classifier ensembles based on binary feature quantisation with application to face recognition , 2008, Appl. Soft Comput..

[10]  Sishaj P. Simon,et al.  A spiking neural network (SNN) forecast engine for short-term electrical load forecasting , 2013, Appl. Soft Comput..

[11]  Zafar A. Khan,et al.  Load forecasting, dynamic pricing and DSM in smart grid: A review , 2016 .

[12]  H. Bevrani,et al.  A fuzzy inference model for short-term load forecasting , 2012, 2012 Second Iranian Conference on Renewable Energy and Distributed Generation.

[13]  Jing Zhao,et al.  Techniques of applying wavelet de-noising into a combined model for short-term load forecasting , 2014 .

[14]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[15]  J. Ben Hadj Slama,et al.  Day-ahead load forecast using random forest and expert input selection , 2015 .

[16]  Farshid Keynia,et al.  A new cascade NN based method to short-term load forecast in deregulated electricity market , 2013 .

[17]  JinXing Che,et al.  Optimal training subset in a support vector regression electric load forecasting model , 2012, Appl. Soft Comput..

[18]  Vladimir Ceperic,et al.  A Strategy for Short-Term Load Forecasting by Support Vector Regression Machines , 2013, IEEE Transactions on Power Systems.

[19]  V.H. Hinojosa,et al.  Short-Term Load Forecasting Using Fuzzy Inductive Reasoning and Evolutionary Algorithms , 2010, IEEE Transactions on Power Systems.

[20]  Haidar Samet,et al.  A new hybrid Modified Firefly Algorithm and Support Vector Regression model for accurate Short Term Load Forecasting , 2014, Expert Syst. Appl..

[21]  Grzegorz Dudek,et al.  Short-Term Load Forecasting Using Random Forests , 2014, IEEE Conf. on Intelligent Systems.

[22]  Jian Ye,et al.  A new ARMAX model based on evolutionary algorithm and particle swarm optimization for short-term load forecasting , 2008 .

[23]  S. A. Soliman,et al.  Short-term electric load forecasting based on Kalman filtering algorithm with moving window weather and load model , 2004 .

[24]  Jaime Lloret,et al.  Artificial Neural Network for Short-Term Load Forecasting in Distribution Systems , 2014 .

[25]  Leo Breiman,et al.  Classification and Regression Trees , 1984 .

[26]  W. R. Christiaanse.  Short-Term Load Forecasting Using General Exponential Smoothing , 1971 .

[27]  Sung-Kwan Joo,et al.  Holiday Load Forecasting Using Fuzzy Polynomial Regression With Weather Feature Selection and Adjustment , 2012, IEEE Transactions on Power Systems.

[28]  Àngela Nebot,et al.  Hybrid methodologies for electricity load forecasting: Entropy-based feature selection with machine learning and soft computing techniques , 2015 .

[29]  Song Li,et al.  Short-term load forecasting by wavelet transform and evolutionary extreme learning machine , 2015 .

[30]  N. Amjady,et al.  Short-Term Bus Load Forecasting of Power Systems by a New Hybrid Method , 2007, IEEE Transactions on Power Systems.

[31]  Hongzhan Nie,et al.  Hybrid of ARIMA and SVMs for Short-Term Load Forecasting , 2012 .